Wide-spread efforts are being made to develop intensity modulated radiotherapy (IMRT) using averse treatment planning, computer-controlled linear accelerators and dynamic multi-leaf collimators (MLC). Although the use of inverse-planning methods may result in computed dose distributions that are highly conformed to the treatment volume, there are potential problems with the accuracy of delivery of the computed dose distribution of the patient, especially for intensity-modulated beams formed by dynamic MLCs and for smaller fields formed by micro MLCs (MMLC). We hypothesize that the application of Monte Carlo tools to the planning and delivery of IMRT will provide significant improvements in the accuracy of IMRT delivery over the accuracy that can be achieved using conventional tools which will lead to an increased theoretical local control with restricted complications. The specific aims of the project are: (1) Simulate dynamic MLC beam delivery. We will show that beamlets and modulated beam profiles calculated with Monte Carlo simulations are more accurate that those computed using current methods in a heterogeneous medium. We will account for the effects of accelerator head scatter, leaf-leakage and photon-backscatter into the monitor chamber associated with the collimator movement and leaf motion. The beamlet and modulated beam distribution will include the effect of phantom heterogeneity, organ motion and patient setup uncertainties. (2) Implement Monte Carlo for inverse planning. We will show that IMRT plans computed using the more accurate beamlets and modulated beam profiles produce accurate dose distributions than those computer by algorithms. We will integrate the more accurate beamlet to improve the efficiency of Monte Carlo simulations for practical routine IMRT applications. (3) Study the impact of Monte Carlo dose calculation. We will show the impact of the improved accuracy on theoretical measures of quality of dose distributions. We will compare the compare the dose distributions calculated by Monte Carlo and conventional algorithms. Effects produced by dose calculation uncertainty, organ motion and patient setup on IMRT treatment will be measured using dose volume histograms (DVH), homogeneity indices (HI), conformity indices (CI) and complication-free control rate (CFCR). The achievement of these aims will provide the radiation oncologists the assurance the IMRT dose distributions are not only desirable but also accurately delivered. The Monte Carlo dose calculation algorithm and the modular inverse-planning system developed in this work will be available for widespread use and implementation, with minimal adjustments, for the bulk for the bulk of the MLCs and MMLCs used for conformal radiotherapy.